Delving deeper into what segmentation is

The difficulty of segmentation is finding actionable business outcomes. It is not difficult to find groups of customers with similar characteristics, but it is difficult to find groups of customers with similar characteristics that are also meaningful.

In computer science, you have the “no free lunch theorem,” which states that “for both static and time-dependent optimization problems, the average performance of any pair of algorithms across all possible problems is the same.” In other words, there is no algorithm that is better than any other algorithm in all situations. This is true for segmentation. There is no segmentation algorithm that is better than any other segmentation algorithm in all situations. This means you need to evaluate the results of your segmentation to see if they are meaningful, but you cannot do so a priori.

You can break segmentation into two categories: clustering and classification...